qcd practice - kiashome.kias.re.kr/mkg/upload/ypschool/jungillec.pdf · 2015. 12. 14. · the...
Post on 17-Feb-2021
5 Views
Preview:
TRANSCRIPT
-
QCD PRACTICE
JUNGIL LEE
Department of Physics
Korea University
Seoul, 136-701
jungil@korea.ac.kr
December 7, 2015
-
2
-
Contents
1 QED Review 1
1.1 Maxwell’s equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Gauge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Classical charged-particle Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 Gauge Invariance and Covariant Derivative . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6 Scalar particle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.7 Time-dependent perturbation theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.8 Propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.9 Photon propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.10 Feynman rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.11 Dirac equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.12 Spinor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.13 Fermion propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2 QCD Lagrangian 31
2.1 QED Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 QCD Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3 Gauge Fixing in QED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.4 Gauge-Fixing and Ghost Terms in QCD . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.5 SU(Nc) algebra summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.6 QCD Feynman rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.6.1 Gluon Vertices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.6.2 Ghost Vertices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3 SU(N) 39
3.1 Generators and structure constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Derivation of completeness relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3 Useful trace formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.4 Adjoint representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.5 SU(3) Clebsch-Gordan coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3
-
4 CONTENTS
4 Tree-Level Calculation 494.1 Cross section formula Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494.2 e+e− → µ+µ− . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504.3 e+e− → q + q̄ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.4 qq̄ → g∗ → q′ + q̄′ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5 Higher Order Correction 555.1 Fermion Self Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.2 Gluon Self Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585.3 Scalar Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.4 Vertex Correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605.5 Gloun Self Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.5.1 Scalar Integrals-Massless fermion loop . . . . . . . . . . . . . . . . . . . . . . . 615.5.2 Scalar Integrals-Massive fermion loop . . . . . . . . . . . . . . . . . . . . . . . 635.5.3 Sign Convention of Π . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.5.4 Fermion Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.5.5 Ghost Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5.6 Gluon Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5.7 Gluon + Ghost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5.8 Gluon Self Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.6 ZQ Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.7 Zg Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.8 1-gluon Exchange Vertex Contribution λ1 . . . . . . . . . . . . . . . . . . . . . . . . . 685.9 tri-gluon Vertex Contribution λ2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695.10 Sum of vertex corrections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695.11 Charge renoramlization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695.12 UV/IR-finite final result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6 Dimensional Regularization 736.1 Minkowski space volume integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736.2 Wick Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746.3 Useful Integrals in the Eucledian Space . . . . . . . . . . . . . . . . . . . . . . . . . . . 776.4 Feynman Parametrization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776.5 Γ function characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7 One Loop Recurrence Relations 797.1 Basic Scalar Integral Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 797.2 Feynman Trick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807.3 primitive recurrence relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807.4 Integration by parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817.5 In for non-integer n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
-
Chapter 1
QED Review
1.1 Maxwell’s equations
1. Let us define covariant(xµ) and contravariant (xµ) vectors as
xµ ≡ (t,+x) , xµ ≡ (t,−x) . (1.1)
2. Show that
∂µ ≡∂
∂xµ=
(∂
∂t,+∇
), ∂µ ≡ ∂
∂xµ=
(∂
∂t,−∇
)(1.2a)
3. Show that ∂µ(∂µ) transforms like xµ(x
µ).
4. Show
∂2 =∂
∂t
2
−∇2. (1.3)
5. Show
∂ · J = ∂∂tJ0 +∇ · J . (1.4)
6. Using the potential Aµ = (φ,A), express the electromagnetic fields
E = −∂A∂t−∇φ, (1.5a)
B = ∇×A. (1.5b)
7. Show that
Ei = − ∂∂x0
Ai − ∂∂xi
A0 = −∂0Ai + ∂iA0, (1.6a)
Bi = �ijk∂
∂xjAk = −�ijk∂jAk. (1.6b)
1
-
8. Defining the field-strength tensor
Fµν ≡ ∂µAν − ∂νAµ, (1.7)
show that
Ei = ∂iA0 − ∂0Ai = −F 0i,
Bi = −12�ijkF jk.
9. Express the field strength tensor in a matrix form as
Fµν =
0 −E1 −E2 −E3E1 0 −B3 B2E2 B3 0 −B1E3 −B2 B1 0
, (1.8a)
Fµν = gµαFαβgβν =
0 E1 E2 E3
−E1 0 −B3 B2−E2 B3 0 −B1−E3 −B2 B1 0
. (1.8b)10. Show that
Fµν = Fµν(E → −E). (1.9)
11. Let us define the dual field-strength tensor Fµν as
Fµν = −12�µναβFαβ, (1.10)
where �µναβ is a totally antisymmetric tensor and �0123 = −�0123 = +1.
12. Show that
Fµν =
0 −B1 −B2 −B3B1 0 E3 −E2B2 −E3 0 E1B3 E2 −E1 0
= Fµν(E → B,B → −E). (1.11)13. Show that
∇ · V = ∂iV i,(∂
∂tV
)i= ∂0V i (1.12a)
(∇× V )i = �ijk∂jV k = −�ijk∂jV k (1.12b)�ijk∂i∂jV k = 0 (1.12c)
�ijk∂iF jk = 0 (1.12d)
(∇×E)i = 12�ijk∂0F jk,
∂Bi
∂t= −1
2�ijk∂0F jk (1.12e)
(∇×B)i = ∂jF ji, −∂Ei
∂t= ∂0F
0i (1.12f)
2
-
14. Show that
∇ ·E = ρ → ∂0F 00 + ∂iF i0 = J0 (1.13a)
∇×B − ∂E∂t
= J → ∂0F 0i + ∂jF ji = J i (1.13b)
∇ ·B = 0 → 12�ijk∂iF jk ≡ 0 : ∂iF i0 = 0 (1.13c)
∇×E + ∂B∂t
= 0 → 12�ijk∂0F jk − 1
2�ijk∂0F jk = 0 (1.13d)
: ∂0F0i + ∂jF ji = 0 (1.13e)
15. You have shown that Maxwell’s equations reduce into the form
∂µFµν = Jν , ∂µFµν = 0 (1.14)
1.2 Gauge
16. Show that ∂µ∂νχ− ∂ν∂µχ = 0, where χ is a Lorentz scalar function.
17. Show that Fµν is invariant under the gauge transformation
Aµ → Aµ + ∂µχ, (1.15)
and the electromagnetic fields E and B are also gauge invariant.
18. Show that the Maxwell’s equations are gauge invariant.
19. Let us use the Lorentz gauge ∂ ·A = 0. Show that the Maxwell’s equations reduce into
∂2Aµ = Jµ. (1.16)
20. We can make another gauge transformation
Aµ → Aµ + ∂µΛ. (1.17)
Show that Λ satisfies the Lorentz condition ∂2Λ = 0.
21. Show thatAµ = �µ(p)e−ip·x. (1.18)
is a solution to the free photon(Jµ = 0) equation with
p2 = 0, � · p = 0. (1.19)
22. Choosing Λ = iae−ip·x, where a is a scalar, show that the gauge invariance condition ensuresthat the transformation
�µ → �′µ = �µ + apµ (1.20)
does not change physical results.
3
-
23. If we choose a such that �′0 = 0, show that
� · p = 0 (1.21)
and it is equivalent to the Coulomb gauge ∇ ·A = 0.
24. Show that the transverse condition � · p = 0 and Coulomb gauge condition � · p = 0 restricts thephoton to have only two degrees of freedom
�µ = (0, 1, 0, 0), (0, 0, 1, 0) (1.22)
if pµ = (E, 0, 0, E).
25. We have shown that there are only two independent parameters describing the polarization vector�µ of the photon. There are only two polarization(spin) states for a massless spin-1 particle.
26. Show that spin-1 wavefunction is expressed in terms of spherical harmonics for |` = 1, `z〉 statesas
Y ±1 (θ, φ) = ∓√
3
8πsin θe±iφ =
√3
4πr̂ · �(±), (1.23a)
Y 01 (θ, φ) =
√3
4πcos θ =
√3
4πr̂ · �(0). (1.23b)
where r̂ = (sin θ cosφ, sin θ sinφ, cosφ).
27. We can express the spherical harmonics in Cartesian coordinate system, which is convenient forvector transformation. The polarization vectors �(λ) are defined by
�(±) = ∓ 1√2
(x̂± iŷ) = ∓ 1√2
(1,±i, 0) (1.24a)
�(0) = ẑ = (0, 0, 1) (1.24b)
28. Show that the polarization vectors are expressed using four-vector nontation as
�µ(±) = ∓ 1√2
(0, 1,±i, 0), (1.25a)
�µ(0) = (0, 0, 0, 1). (1.25b)
29. Show that massless photon’s wavefunction is a linear combination of �µ(±). The longitudinalstate �µ(0) is not allowed.
30. Show the orthogonality conditions
�∗(λ) · �(λ′) = δλλ′ , (1.26a)�∗(λ) · �(λ′) = −δλλ′ . (1.26b)
31. Show that the spin sum of the polarization tensor is
∑λ=±
�i∗(λ)�j(λ) = δij⊥ = δij − ẑiẑj =
1 0 00 1 00 0 0
ij
, (1.27a)
if kµ = (k, 0, 0, k).
4
-
32. Show that the spin sum of the polarization tensor is
∑λ=±
�i∗(λ)�j(λ) = δij⊥ = δij − k
ikj
k2. (1.28a)
if kµ = (|k|,k). The polarization sum is for the Coulomb gauge, where ∇ ·A = 0.
33. Defining n = (1, 0, 0,−1) when p = (E, 0, 0, E), show that we may write it in a Lorentz covariantform
∑λ=±
�∗µ(λ)�ν(λ) =
0 0 0 00 1 0 00 0 1 00 0 0 0
µν
= −gµν + pµnν + nµpν
p · n(1.29a)
34. The formula
Πµν =∑λ=±
�∗µ(λ)�ν(λ) = −gµν + pµnν + nµpν
p · n(1.30)
is valid for any light-like vector n2 = 0 satisfying � ·p 6= 0, n ·p 6= 0, and n · � = 0. Show explicitlythat
Πµµ = −2, (1.31a)nµΠ
µν = 0, nνΠµν = 0, (1.31b)
pµΠµν = 0, pνΠ
µν = 0. (1.31c)
The polarization sum is for the light-cone gauge, where n ·A = 0 with n2 = 0.
35. We can extend our formula to the case n2 6= 0 keeping � · p 6= 0, n · p 6= 0, and n · � = 0. Derive
Πµν =∑λ=±
�∗µ(λ)�ν(λ) = −gµν + pµnν + nµpν
p · n− n
2pµpν
(p · n)2. (1.32)
from the conditions
Πµµ = −2, (1.33a)nµΠ
µν = 0, nνΠµν = 0, (1.33b)
pµΠµν = 0, pνΠ
µν = 0, (1.33c)
The polarization sum is for the axial gauge, where n ·A = 0 with n2 6= 0.
1.3 Lagrangian
36. Euler-Lagrange equation: Action is defined by
S =
∫d4xL(φ, ∂µφ). (1.34)
5
-
Show that under the variation φ→ φ+ δφ, where φ and ∂µφ are fixed at the end points
δS =
∫d4xδL(φ, ∂µφ) =
∫d4x
[∂L∂φ
δφ+∂L
∂(∂µφ)δ∂µφ
](1.35a)
=
∫d4x
[∂L∂φ− ∂µ
∂L∂(∂µφ)
]δφ, (1.35b)
where we neglected the surface term∫d4x∂µ
[∂L
∂(∂µφ)δφ
]=
∫daµ
∂L∂(∂µφ)
δφ. (1.36a)
37. Show that δS = 0 if φ satisfies the Euler-Lagrange equation
∂L∂φ− ∂µ
∂L∂(∂µφ)
= 0. (1.37)
38. real scalar field If
L = 12
(∂µφ∂µφ−m2φ2), (1.38)
show that the Euler-Lagrange equation is the Klein-Gordon equation
(∂2 +m2)φ = 0. (1.39)
This leads to p2 = m2.
39. Symmetry and conserved current If the Lagrangian is invariant under a transformationφ→ φ+ δφ, show that
δL(φ, ∂µφ) =∂L∂φ
δφ+∂L
∂(∂µφ)δ∂µφ, (1.40a)
=
[∂µ
∂L∂(∂µφ)
]δφ+
∂L∂(∂µφ)
∂µ(δφ) (1.40b)
= ∂µ
[∂L
∂(∂µφ)δφ
]= 0. (1.40c)
If there is symmetry, there is a corresponding conserved quantity.
40. The conserved (∂ · J = 0) current
Jµ ∝ ∂L∂(∂µφ)
δφ. (1.41)
41. If the current is vanishing on a boundary surface, charge inside the surface is conserved
∂Q
∂t= −
∫d3x∇ · J = −
∫da · J = 0, (1.42a)
where Q =∫d3xJ0.
6
-
42. Show that
−12FµνFµν = E
2 −B2 (1.43a)
43. Show that
L = −14FµνFµν − JµAµ =
1
2
(E2 −B2
)− ρφ+ J ·A
44. Show that the Euler-Lagrange equation for the Lagrangian is the Maxwell’s equations
∂µFµν = Jν ; (∂2 − ∂µ∂ν)Aν = Jν (1.44)
1.4 Classical charged-particle Lagrangian
45. Consider a particle with the charge q and mass m moving in an external electromagnetic field.The equation of motion is
dE
dt= qv ·E, dp
dt= q (E + v ×B) . (1.45a)
46. Let us define the four-velocity in Lorentz covariant notation
uµ =dxµ
dτ= γ(1,v). (1.46)
show that u2 = 1.
47. Using the four-velocity in Lorentz covariant notation, show that
dpµ
dτ= qγ (v ·E,E + v ×B) = qFµνuν , (1.47)
where pµ = muµ and m is the rest mass of the particle.
48. Show that
Fµνuν = γ
0 −E1 −E2 −E3E1 0 −B3 B2E2 B3 0 −B1E3 −B2 B1 0
1−v1−v2−v3
(1.48a)
= γ
v1E1 + v2E2 + v3E3
E1 + (v2B3 − v3B2)E2 + (v3B1 − v1B3)E3 + (v1B2 − v2B1)
µ
. (1.48b)
7
-
49. Free particle Lagrangian: Consider a free particle with a mass m. The action must be Lorentzinvariant and the only available Lorentz invariant scalar is m.
S =
∫ tfti
Ldt =
∫ tfti
γLdτ = f(m). (1.49)
and Lγ must be a scalar. And the dimension must be order of mass. Let us try
L = −mγ
= −m√
1− v2. (1.50)
Show that the Euler-Lagrange equation is
∂
∂t
(∂L
∂vi
)− ∂L∂xi
= 0, (1.51a)
∂L
∂vi=
mvi√1− v2
= γmvi, (1.51b)
∂
∂t(mγvi) = 0. (1.51c)
Momentum is conserved! ← free particle.
50. Show that corresponding Hamiltonian is
pi =∂L
∂vi= − ∂
∂vi
√1− v2 = mv
i
√1− v2
, (1.52a)
H = pivi − L = mv2
√1− v2
+m√
1− v2 (1.52b)
=m√
1− v2− m(1− v
2)√1− v2
+m√
1− v2 (1.52c)
=m√
1− v2= γm =
√p2 +m2. (1.52d)
51. Charged particle Lagrangian: In nonrelativistic quantum mechanics, Lint = −Vint. If weconsider the electrostatic potential in nonrelativistic quantum mechanics,
Lint = −Vint = −qφ. (1.53)
52. Let us construct a Lorentz scalar. As we did for a charged particle, γL must be Lorentz invariantand in the nonrelativistic limit the Lagrangian must reduce the form shown above.
Lint = −qu ·Aγ
, (1.54)
where uµ = γ(1,v) and Aµ = (φ,A).
53. Therefore,
L = −m+ qv ·Aγ
= −m√
1− v2 − qφ+ qv ·A (1.55)
8
-
Conjugate momentum is
P i =∂L
∂vi= γmvi + qAi = (p+ qA)i → v = P − qA
γm, (1.56a)
1 = γ2(1− v2) = γ2 − (P − qA)2
m2, (1.56b)
→ γm =√
(P − qA)2 +m2 → v = P − qA√(P − qA)2 +m2
. (1.56c)
54. Show that
(P − qA) · v = (P − qA)2√
(P − qA)2 +m2, (1.57a)
m
γ=
m2
γm=
m2√(P − qA)2 +m2
. (1.57b)
Then the Hamiltonian is
H = P · v − L = (P − qA) · v + mγ
+ qφ (1.58a)
=
√(P − qA)2 +m2 + qφ (1.58b)
55. The Hamiltonian for a charged particle
H =
√(P − qA)2 +m2 + qφ. (1.59)
is same as that for a free particleH =
√p2 +m2 (1.60)
when we substitute
H → H − qφ (1.61a)p → P − qA. (1.61b)
56. Show that mass-shell condition still holds
p2 = m2, pµ = (E − qφ,P − qA) , (1.62)
where E is the total energy of the particle
57. Nonrelativistic case: Show that in the nonrelativistic(NR) limit
H =
√(P − qA)2 +m2 + qφ→ (P − qA)
2
2m+ qφ. (1.63)
We could have derived the form from the free particle equation
H =p2
2m← [H → H − qφ, p→ P − qA] . (1.64)
9
-
58. Schrödinger equation for a charged particle: For a free particle, we replace
p→ −i∇, H → i ∂∂t, (1.65)
which generate p and E once they act on the free-particle wavefunction e−ip·x, and apply theoperator to the wavefunction ψ.
i∂
∂tψ =
(−i∇)2
2mψ (1.66)
59. Show that the Schrödinger equation for a charged particle is
(H − qφ)ψ = (P − qA)2
2mψ, (1.67a)
i
(∂
∂t+ iqA0
)ψ = −(−∇+ iqA)
2
2mψ. (1.67b)
1.5 Gauge Invariance and Covariant Derivative
60. Let us define the covariant derivative
Dµ ≡ ∂µ + iqAµ =(D0,−D
)=
(∂
∂t+ iqA0,−∇+ iqA
). (1.68)
Show that the shrödinger equation becomes
iD0ψ = −D2
2mψ. (1.69)
61. Let us replace ψ → ψ′ = Uψ, where the transformation keeps the probability
ψ†ψ = ψ′†ψ′ → U †U = 1. (1.70)
Therefore U is unitary.
62. We know physical observables are invariant under the gauge transformation
Aµ → A′µ = Aµ + ∂µχ. (1.71)
We will find there IS a gauge transformation that keeps the Shrödinger equation invariant underthe unitary transformation ψ → ψ′ = Uψ.
D′µ = ∂µ + iqA′µ = ∂µ + iq(Aµ + ∂µχ) = Dµ + iq∂µχ (1.72a)
D′0 = ∂0 + iq(A0 + ∂0χ) (1.72b)
−D′ = −∇+ iq(A−∇χ). (1.72c)
Show, if ∂µχ = iq (∂µU)U †, that
D′0Uψ = UD0ψ, D′Uψ = UDψ (1.73a)
(D′)2Uψ = D′(UDψ) = UD2ψ (1.73b)
Therefore, Schrödinger equation is invariant under the gauge transformation
Aµ → A′µ = UAµU † + iq
(∂µU)U †, ψ → ψ′ = Uψ, U †U = 1 (1.74)
10
-
1.6 Scalar particle
63. Scalar particle and EM interaction Remember
H =√p2 +m2 → H2 = p2 +m2. (1.75)
Klein-Gordon equation is the wave equation for a scalar particle that can be obtained by thereplacement H → i∂0 and p = −i∇, that is pµ → i∂µ
(∂2 +m2)φ = 0. (1.76)
We have checked that the Lagrangian for the equation of motion is
L = 12
(∂µφ∂µφ−m2φ2). (1.77)
If we introduce a complex scalar field, that can be constructed as a linear combination of tworeal scalar fields as
φ =1√2
(φ1 + iφ2) (1.78)
φ∗ 6= φ.
64. Using the fact that φ1 and φ2 are satisfying the Klein-Gordon equation, show that φ and φ∗ also
satisfies the Klein-Gordon equation.
(∂2 +m2)φ = 0, (∂2 +m2)φ∗ = 0. (1.79)
65. Show that the Lagrangian is
L = ∂µφ∗∂µφ−m2φ∗φ. (1.80)
Now we can introduce the covariant derivative to the complex scalar field.
L = (Dµφ)†(Dµφ)−m2φ†φ, Dµ = ∂µ + iqAµ. (1.81)
66. Show that the Lagrangian is invariant under the gauge transformation
φ → Uφ, U †U = 1, (1.82a)
Aµ → UAµU † + iq
(∂µU)U †. (1.82b)
67. The Lagrangian for the electromagnetic field is
L = −14FµνFµν − JµAµ. (1.83a)
Show that the current Jµ is
Jµ = +iq[φ†∂µφ− (∂µφ†)φ
], (1.84)
by expanding the covariant derivative.
11
-
68. Show that the Jµ is conserved∂µJ
µ = 0 (1.85)
by using the Klein-Gordon equation.
69. Show that
Q =
∫d3xJ0 = 0 (1.86)
for a real scalar field φ∗ = φ.
70. Show that the current becomesJµ = q × |N |22pµ. (1.87)
if we use the free-particle wavefunction φ = Ne−ip·x and φ∗ = N∗eip·x, where N is the normal-ization factor.
71. Show that, if∫d3x = V ,
Q =
∫d3xJ0 = q × |N |22EV. (1.88)
72. Choose the covariant normalization N = 1/√V and show that the charge inside the volume
V =∫d3x is
Q =
∫d3xJ0 = q × 2E. (1.89)
73. q is the charge of the particle φ.
74. Show that 2E is the number of particles in V .
75. Show that the charge Q in V is not Lorentz invariant.
76. The number of particles in V is 2m if the particle is at rest. Explain why the density is increasingwith a factor E/m compared to that for the rest frame because of the length contraction.
77. Negative energy solution to the Klein-Gordon equation: Let us go back to the Klein-Gordon equation
(∂2 +m2)φ = 0. (1.90)
Show that there are two solutions
φ+ = Ne−ip·x, φ− = Ne
+ip·x (1.91)
where p = (E,p) with E =√p2 +m2 > 0.
78. Show that Jµ(φ = φ+) = q × 2pµ|N |2 and Jµ(φ = φ−) = −q × 2pµ|N |2.
79. Show that Jµ(φ = φ−) = q × 2(−pµ) means that a negative energy particle with charge +q isflowing from the future to the past.
80. Show that Jµ(φ = φ−) = (−q) × 2pµ means that a positive energy particle with charge −q isflowing from the past to the future.
81. Combining the two equivalent statements, we conclude as follows. Once we know how to dealwith a scalar particle with charge +q, the wavefunction for a positive-energy scalar particle withcharge −q can be described in terms of the negative-energy scalar particle with charge +q flowingbackward!
12
-
1.7 Time-dependent perturbation theory
82. Consider
H = H0 + V, H0 =p2
2m, (1.92a)
H0φn = Enφn,
∫d3xφ∗m(x)φn(x) = δmn. (1.92b)
Write the wavefunction ψ in terms of the eigenfunctions of the unperturbed Hamiltonian as
ψ =∑n
an(t)φne−iEnt. (1.93)
Solve an satisfying
i∂
∂tψ = (H0 + V )ψ (1.94)
to get
i∑n
∂an(t)
∂tφn(x)e
−iEnt =∑n
an(t)V (t,x)φn(x)e−iEnt. (1.95)
83. Show that an(t→ −∞) = δni means the initial state is monochromatic
ψ(t = −∞) = φi(x)e−iEit. (1.96)
84. Convoluting ∫d3xφ†f (x)e
iEf t, (1.97)
show that
afi = δfi − i∫dtd3xφ†f (x)V (t,x)φi(x)e
i(Ef−Ei)t (1.98)
85. In short
ψ(t→∞) = Sψ(t→ −∞), (1.99a)S = 1 + iT , (1.99b)
iT = −i∫d4xφ∗f (x)V φi(x) = i
∫d4xLint = Lint. (1.99c)
86. Show that, if T is HermitianT † = T ↔ Lint, (1.100)
S is unitarity S†S = 1. Therefore,
ψ†ψ(t =∞) = ψ†ψ(t = −∞). (1.101)
13
-
87. Energy-momentum conservation: If the potential is independent of time V (t,x) = V (x),iTfi becomes
iTfi = −iVfi∫ ∞−∞
dtei(Ef−Ei)t = −iVfi(2π)δ(Ef − Ei), (1.102a)
Vfi =
∫d3xφ†f (x)V (x)φi(x), (1.102b)
and the energy is conserved.p2f2m
=p2i2m
. (1.103)
88. In general, pf 6= pi. Where is the lost momentum, pi − pf? It has been transferred to thepotential because
V (x) =
∫d3k
(2π)3V (k)e+ik·x, (1.104a)
φ†f (x) = Nfe+ipf ·x, (1.104b)
φi(x) = Nie−ipi·x. (1.104c)
89. Show that
Vfi = NfNi
∫d3kV (k)δ(k + pf − pi) = NfNiV (pi − pf ). (1.105)
Therefore, momentum is conserved k + pf = pi.
1.8 Propagator
90. Propagator for a scalar particle: Let us recall the transition matrix
iT = i∫d4xLint. (1.106)
We choose Lint = −φ†J which is analogous to the electromagnetic interaction lagrangian −JµAµ.Resulting lagrangian for a scalar field is
L = φ†(−∂2 −m2
)φ− φ†J, (1.107)
where we neglect the surface term. Then the wave equation becomes
∂L∂φ†
= 0 →(∂2 +m2
)φ = −J. (1.108)
Show that
φ(x) =
∫dy4∆F (x− y)J(y), (1.109)
if (∂2 +m2
)∆F (x) = −δ(x). (1.110)
Hint: Act ∂2 +m2 to both sides.
14
-
91. Show that
iT = i∫d4xLint = −i
∫d4xJ(x)φ(x) (1.111a)
= −i∫d4xd4yJ(x)∆F (x− y)J(y) (1.111b)
=
∫d4xd4y [−iJ(x)] [i∆F (x− y)] [−iJ(y)] (1.111c)
92. i∆F (x − y) is the Feynman propagator. It describes the propagation of a scalar field from aspace-time point y to x.
93. Show that ∫ ∞−∞
dp0
2π
e−ip0x0
p0 − E + i�= −i e−iEx0 , if x0 > 0, (1.112)
Hint: Use Cauchy integral formula. Explain why the above integral is same as the complexintegral along a contour closed on the lower-half plain.
94. Show that ∫dp0
2π
e−ip0x0
p0 + E − i�= +i eiEx
0, if x0 < 0, (1.113)
Explain why the above integral is same as the complex integral along a contour closed on theupper-half plain.
95. Show that ∫dp0
2π
e−ip0x0
p2 −m2 + i�=
∫dp0
2π
e−ip0x0
(p0)2 − (E2 − i�)
=
∫dp0
2π
e−ip0x0
(p0 + E − i�)(p0 − E + i�)
=−i2E
(θ(x0)e−iEx
0+ θ(−x0)e+iEx0
)(1.114a)
where E =√p2 +m2 > 0, �→ 0+ and
θ(x) =
{1 if x > 00 if x < 0
(1.115)
96. Explain why the sign of i� is important.
97. Explain what is wrong if � is finite.
98. Show that ∫d4p
(2π)4i
p2 −m2 + i�e−ip·x (1.116a)
=
∫d3p
2E(2π)3
[θ(x0)e−i(Et−p·x) + θ(−x0)ei(Et−p·x)
](1.116b)
=
∫d3p
2E(2π)3[θ(x0)e−ip·x + θ(−x0)eip·x
](1.116c)
where E =√m2 + p2 > 0 and p0 = E on the last line.
15
-
99. Show that the retarded term θ(x0) is for the particle propagating to the future.
100. Show that the advanced term θ(−x0) term is for the particle to the past.
101. Show that
∆F (x) =
∫d4p
(2π)4e−ip·x
p2 −m2 + i�(1.117)
is the solution to the equation (∂2 +m2
)∆F (x) = −δ(x) (1.118)
1.9 Photon propagator
102. The iT matrix for a photon field is
iT = i∫d4xLint = −i
∫d4xJµAµ, (1.119)
where the equation of motion is Maxwell’s equation
∂νFνµ = Jµ;
(∂2gµν − ∂µ∂ν
)Aν = J
µ. (1.120)
103. We want to find the solution to(∂2gµα − ∂µ∂α
)(DF )αν(x) = g
µνδ(x). (1.121)
Aµ(x) =
∫dy4DµνF (x− y)Jν(y) (1.122a)
if(∂2gµα − ∂µ∂α
)(DF )αν = g
µνδ(x). (1.122b)
104. Show that
T = i∫d4xLint = −i
∫d4xJµ(x)A
µ(x) (1.123)
= −i∫d4xd4yJµ(x)D
µνF (x− y)Jν(y) (1.124)
=
∫d4xd4y [−iJµ(x)]
[iDµνF (x− y)
][−iJν(y)] (1.125)
i∆µνF (x − y) is the Feynman propagator. It describes the propagation of a vector field from aspace-time point y to x.
105. Propagator in the Feynman gauge: If we choose the Lorentz gauge, ∂ ·A = 0, and we mayneglect the term ∂µ∂ν terms in the wave equations. Maxwell’s equation becomes
∂2Aµ = Jµ. (1.126)
16
-
Show that
DµνF (x) =
∫d4p
(2π)4−gµν
p2 + i�e−ip·x = −gµν∆F (x) with m = 0 (1.127)
is the solution to the equation
∂2gµα(DF )αν = gµνδ(x). (1.128)
106. Show that the wave equation∂2Aµ = Jµ (1.129)
is actually from the Lagrangian
L = −14FµνFµν − JµAµ −
1
2α(∂ ·A)2 (1.130)
where α = 1. The term added to the original Lagrangian is the gauge-fixing term.
107. The photon propagatorDµνF (x) = −g
µν∆F (x) (1.131)
is defined in the Feynman gauge, a special case of the Lorentz gauge.
108. Propagator in the Lorentz gauge Show that the equation of motion for the photon field inthe Lagrangian
L = −14FµνFµν − JµAµ −
1
2α(∂ ·A)2 (1.132)
is [∂2gµα +
(1
α− 1)∂µ∂α
]Aα = J
µ (1.133)
and the propagator D must satisfy[∂2gµα +
(1
α− 1)∂µ∂α
]Dαν(x) = g
µνδ(x). (1.134)
Note that we are using the Lorentz gauge ∂ ·A = 0.
109. Propagator in the Lorentz gauge Show that
Dµν(x) =
∫d4p
(2π)4
−gµν + (1− α) pµpν
p2
p2 + i�e−ip·x (1.135)
=
∫d4p
(2π)4
[−gµν + (1− α) p
µpν
p2
]∆F (p)e
−ip·x (1.136)
∆F (p) =1
p2 + i�(1.137)
is the solution to the equation[∂2gµα +
(1
α− 1)∂µ∂α
]Dαν(x) = g
µνδ(x) (1.138)
17
-
110. Propagator in the axial gauge We can choose the axial gauge n ·A = 0. In this case, gaugefixing term can be written in the form to have our Lagrangian
L = −14FµνFµν − JµAµ −
1
2α(n ·A)2 . (1.139)
111. Show that
iDµν(x) = i
∫d4p
(2π)4
[−gµν + n
µpν + pµnν
n · p− (n2 + αp2) p
µpν
(n · p)2
]∆F (p)e
−ip·x (1.140)
∆F (p) =1
p2 + i�(1.141)
is the gluon propagator in the axial gauge.
112. Massive spin-1 propagator The Lagrangian for a massive spin-1 field is just like that for thephoton except that the particle has nonvanishing mass. We insert the mass term m
2
2 BµBµ
L = −14FµνFµν +
m2
2BµBµ − JµBµ (1.142)
Fµν = ∂µBν − ∂νBν (1.143)
113. Show that the equation of motion is[(∂2 +m2)gµν − ∂µ∂ν
]Bν = J
µ. (1.144)
114. The T matrix for a massive spin-1 field is
iT = i∫d4xLint = −i
∫d4xJµBµ, (1.145)
where the equation of motion is[(∂2 +m2)gµν − ∂µ∂ν
]Bν = J
µ. (1.146)
115. We want to find the solution to[(∂2 +m2)gµα − ∂µ∂α
]Dαν(x) = g
µνδ(x). (1.147)
Bµ(x) =
∫dy4Dµν(x− y)Jν(y) (1.148)
if[(∂2 +m2)gµν − ∂µ∂ν
]Dαν(x) = g
µνδ(x) (1.149)
Show that
iT = i∫d4xLint = −i
∫d4xJµ(x)B
µ(x) (1.150)
= −i∫d4xd4yJµ(x)D
µν(x− y)Jν(y) (1.151)
=
∫d4xd4y [−iJµ(x)] [iDµν(x− y)] [−iJν(y)] (1.152)
18
-
116. Show that
Dµν(x) =
∫d4p
(2π)4
−gµν + pµpν
m2
p2 −m2 + i�e−ip·x (1.153)
=
∫d4p
(2π)4
[−gµν + p
µpν
m2
]∆F (p)e
−ip·x (1.154)
∆F (p) =1
p2 −m2 + i�(1.155)
is the solution to the equation[(∂2 +m2)gµν − ∂µ∂ν
]Dαν(x) = g
µνδ(x) (1.156)
Note that p2 6= m2 in general.
1.10 Feynman rules
Momentum-space Feynman rule Let us go back to the T matrix
iT = i∫d4xLint (1.157)
=
∫d4xd4y [−iJ(x)] · [iD(x− y)] · [−iJ(y)] (1.158)
where indices are suppressed for vector particles.
117. Expand the currents as
J(y) =
∫d4k
(2π)4J(k)e−ik·y (1.159)
and show that
iT = i∫d4xLint (1.160)
=
∫d4xd4y [−iJ(x)] · [iD(x− y)] · [−iJ(y)] (1.161)
=
∫d4p
(2π)4[−iJ(−p)] · [iD(p)] · [−iJ(p)] . (1.162)
118. Consider monochromatic currents
J(x) = J(p1) = N2Ĵ(p1)e
−ip1·x (1.163)
J(y) = J(p2) = N2Ĵ(p2)e
−ip2·y (1.164)
where N2 came from the normalization of the initial and final states involving the current.Show that in this case
iT = i∫d4xLint = N4(2π)4δ4(p1 + p2)M (1.165)
M =[−iĴ(p2)
]· [iD(p = p1 = −p2)] ·
[−iĴ(p1)
](1.166)
19
-
119. Show that the propagators for a scalar field and massless/massive vector fields are even functions;
D(x) = D(−x), D(p) = D(−p) (1.167)
Hint: Look into the partial differential equation for the propagator.
120. Show that
|T |2 =∣∣∣∣i∫ d4xLint∣∣∣∣2 = N8 [(2π)4δ4(p1 + p2)]2 |M|2 (1.168)
= N8 × V T × (2π)4δ4(p1 + p2) |M|2 (1.169)
M =[−iĴ(p2)
]· [iD(p = ±p1 or ± p2)] ·
[−iĴ(p1)
](1.170)
Hint: (2π)4δ4(0) =∫dtd3x.
121. Show that the probability of the transition per unit volume per unit time is
P =|T |2
TV=
1
V 4× (2π)4δ4(p1 + p2) |M|2 (1.171)
M =[−iĴ(p2)
]· [iD(p = ±p1 or ± p2)] ·
[−iĴ(p1)
](1.172)
where we used N = 1/√V .
122. Cross section Show that the number density of a particle with energy E in V is
2E
V. (1.173)
123. Show that the flux of the two colliding particle is
F = |va − vb| ×2EaV× 2Eb
V(1.174)
= 4 (|pa|Eb + |pb|Ea) /V 2 (1.175)
= 4√
(pa · pb)2 −m2am2b/V2 (1.176)
where vi, pi, and Ei are the velocity, momentum and energy of the i−th particle.
124. Show that the number of states in the final state is
dNfinal states = V θ(p01)δ(p
21 −m21)
d4p1(2π)4
V θ(p02)δ(p22 −m22)
d4p2(2π)4
=V d3p1
2E1(2π)3V d3p2
2E2(2π)3(1.177)
if there are two final particles.
20
-
125. Cross section is defined by
dσ =P
F× dNfinal states
=1
V 4(2π)4δ(pa + pb − p1 − p2)|M|2
× V2
4√
(pa · pb)2 −m2am2b× V d
3p12E1(2π)3
V d3p22E2(2π)3
=|M|2
4√
(pa · pb)2 −m2am2b(2π)4δ(pa + pb − p1 − p2)
d3p12E1(2π)3
d3p22E2(2π)3
. (1.178)
126. We usually define the phase space dΦ after including the energy-mpmentum delta function
dΦ = (2π)4δ
(P −
∑i
pi
)dNfinal states, (1.179)
where P is the sum of initial momenta pi is the momentum of the i−th final-state particle.
127. We find V dependence exactly cancels. If we redefine the phase space
dΦ = (2π)4δ(pa + pb − p1 − p2)n∏i=1
d3pi2Ei(2π)3
(1.180)
we have
dσ =|M|2
4√
(pa · pb)2 −m2am2bdΦ (1.181)
128. Show that the mass term 12m2AµAµ for a gauge field is NOT invariant under gauge transforma-
tion
Aµ → UAµU † + iq
(∂µU)U †, ψ → Uψ, Dµ = ∂µ + iqAµ (1.182)
This guarantees that the gauge field is massless.
129. Show that gauge field is travelling with the speed of light.
1.11 Dirac equation
130. We would like to construct a relativistically covariant theory for a fermion. If we ignore thespin, the equation must reduce to the Klein-Gordon equation. But we want to have an equationwhich has linear time derivative instead of ∂/∂t2, which appears in the Klein-Gordon equation.
i∂ψ
∂t= Hψ, H = −iα ·∇+ βm (1.183)
21
-
131. Show that ψ must include the four states
|1〉 = | ↑, E > 0〉, |2〉 = | ↓, E > 0〉, (1.184)|3〉 = | ↑, E < 0〉, |4〉 = | ↓, E < 0〉 (1.185)
132. Show that H is Hermitian.
133. Show that
〈m|H|n〉 =√p2 +m2, if m = n = 1, 2 (1.186)
〈m|H|n〉 = −√p2 +m2, if m = n = 3, 4 (1.187)
〈m|H|n〉 = 0, if m 6= n (1.188)∑n
〈n|H|n〉 = 0 (1.189)
134. Show that α1, α2, α3, and β are Hermitian.
135. If the Dirac equation is equivalent to the Klein-Gordon equation if we neglect the spin dependenceand the sign of the energy, show
H2 = −∇2 +m2 → (∂2 +m2)ψ = 0 (1.190)
136. Show that the condition requires
1
2
(αiαj + αjαi
)= δij (1.191)
αiβ + βαi = 0 (1.192)
β2 = 1 (1.193)
137. Show that the conditions (αi)2 = 1 and β2 = 1 require that the eigenvalues for the matirces are±1.From now on we will use the notation
{A,B} ≡ AB +BA. (1.194)
138. Let us choose the basis so that the first two components are positive-energy components andthe other two are negative-energy components. By taking p = 0, show that
β =
1 0 0 00 1 0 00 0 −1 00 0 0 −1
→ ( 1 00 −1)
(1.195)
139. From now on we express any 4× 4 matrix in spinor space in terms of 2× 2 block matrices.
140. Show that β2 = 1.
22
-
141. Let us try arbitrary 4× 4 matrices
αi =
(ai bi
ci di
)(1.196)
142. Show that the condition {β, αi} = 0 requires
αi =
(0 bi
ci 0
)(1.197)
143. Show that the condition α† = α requires c = b†.
144. Show that the condition {αi, αj} = 2δij requires
bi(bj)† + bj(bi)† = 2δij (1.198)
(bi)†bj + (bj)†bi = 2δij (1.199)
145. Choose the term i = j to find bi is unitary
bi(bi)† = (bi)†bi = 1 (1.200)
The solution is the 3 = 22 − 1 SU(2) generators, Pauli matrices; bi = σi, where Pauli matricesare
σ1 =
(0 11 0
), σ2 =
(0 −ii 0
), σ3 =
(1 00 −1
)(1.201)
Therefore, we have
β =
(1 00 −1
), α =
(0 σσ 0
)(1.202)
146. Show that
σiσj = δij + i�ijkσk (1.203)
{σi, σj} = 2δij (1.204)[σi, σj
]≡ σiσj − σjσi = +2i�ijkσk (1.205)
147. Show that
a · σ b · σ = a · b+ ia× b · σ (1.206)
148. Show that
a ·α b ·α = a · b+ ia× b ·Σ, Σ =(σ 00 σ
)(1.207)
23
-
149. Defining γµ = (γ0,γ) = (β, βα), show that
γ0 =
(1 00 −1
), γ =
(0 σ−σ 0
)(1.208)
150. Show that the anticommutation relations for β and αi’s reduce to a single formula
{γµ, γν} = 2gµν (1.209)
where the 4× 4 identity matrix
151. Let us define
/a ≡ γµaµ (1.210)
Show that
/a/b+ /b/a = 2a · b, /a2 = a2 (1.211)
152. Show that(γµ)† = γµ, (γµ)
† = γµ (1.212)
153. Show thatγ0(γµ)†γ0 = γµ, γ0(γµ)
†γ0 = γµ (1.213)
Multiply β to the original Dirac equation and find
(iγµ∂µ −m)ψ = 0 (1.214)
Taking the Hermitian adjoint and find
−i∂µψ̄γµ − ψ̄m = 0 (1.215)
where ψ̄ = ψ†γ0.
154. Show that the Euler-Lagrange equation for the ψ field in the Lagrangian
L = ψ̄(iγµ∂µ −m)ψ (1.216)
gives the Dirac equation(iγµ∂µ −m)ψ = 0 (1.217)
155. Show that the Lagrangian can also be written as
L = −i(∂µψ̄
)γµψ −mψ̄ψ (1.218)
if we neglect the surface term.
156. Show that the Euler-Lagrange equation for ψ† field is
−i∂µψ̄γµ − ψ̄m = 0 (1.219)
24
-
157. Show that the Lagrangian for the Quantum electrodynamics
L = ψ̄(iγµDµ −m)ψ −1
4FµνFµν (1.220)
= −i (Dµψ)† γ0γµψ −mψ̄ψ −1
4FµνFµν (1.221)
is invariant under the gauge transformation
ψ → Uψ, Aµ → UAµU † + iq
(∂µU)U †, Dµ = ∂µ + iqAµ (1.222)
158. Introducing the covariant derivative Dµ = ∂µ + iqAµ, show that
L = ψ̄(iγµ∂µ −m)ψ − JµAµ (1.223)Jµ = qψ̄γµψ (1.224)
159. Making use of the Dirac equations
iγµ∂µψ = mψ, − i(∂µψ̄
)γµ = mψ̄ (1.225)
show that the current is conserved
∂µJµ = 0 (1.226)
1.12 Spinor
160. Let us consider an electron with momentum pµ and z−compenent spin s, where p0 = E > 0.Show that the wavefunction ψ(x) = u(p, s)e−ip·x should be normalized to be
ψ†(x)ψ(x) = 2E → u†(p, s)u(p, s′) = 2Eδss′ . (1.227)
161. Using the length contraction, show that the number of electron in the whole space of volume Vis 1.
162. Show thatū(p)u(p) = 2m (1.228)
using u†(p, s)u(p, s′) = 2mδss′ and Lorentz covariance only.
163. Remind the fact that ψ̄γµψ is transforming like a four vector. Using the result u†(p, s)u(p, s) =2E and Lorentz covariance only, show that
ū(p)γµu(p) = 2pµ. (1.229)
164. Using spinors for an electron at rest p = (m,0), show that∑s
u(p, s)ū(p, s) = 2m×(
1 00 0
)= 2m× 1
2(1 + γ0). (1.230)
25
-
165. Using Lorentz covariance, show that∑s
u(p, s)ū(p, s) = 2m× p/+m2m
= p/+m. (1.231)
166. Let us consider a negative-energy electron with momentum −pµ and z−compenent spin s, wherep0 = E > 0. Show that the wavefunction ψ(x) = u(−p, s)e+ip·x should be normalized to be
ψ†(x)ψ(x) = 2E → u†(−p, s)u(−p, s′) = 2Eδss′ . (1.232)
167. Explain why it is not proportional to −2E < 0 but proportional to 2E > 0.
168. Using the length contraction, show that the number of electron in the whole space of volume Vis 1.
169. Show that
ū(−p, s)u(−p, s′) = −2mδss′ (1.233)
using u†(−p, s)u(−p, s′) = 2mδss′ and Lorentz covariance only.
170. Show that ∑s
u(−p, s)ū(−p, s) = 2m×(
0 00 −1
)= 2m× 1
2(−1 + γ0). (1.234)
171. Using Lorentz covariance, show that∑s
u(−p, s)ū(−p, s) = 2m× p/−m2m
= p/−m. (1.235)
172. Explain why this formula cannot be obtained if we substitute p→ −p to the postive energy case∑s
u(p, s)ū(p, s) = p/+m. (1.236)
173. Following the results for the charged scalar field, we would like to make use of the negative-energysolution for the positive energy antiparticle with momentum −(−p) = p. Then the wavefunctionfor the antiparticle with momentum p can be v̄(p) ≡ ū(−p). Show that
v†(p, s)v(p, s′) = 2Eδss′ , (1.237a)
v̄(p, s)v(p, s′) = −2mδss′ , (1.237b)∑s
v(p, s)v̄(p, s) = p/−m, (1.237c)
where p = (E,p) and E =√p2 +m2 > 0.
174. Wavefunction for a positive-energy antiparticle with momentum pµ is
v̄(p)e−ip·x. (1.238)
26
-
175. For the final state, we use
v(p)e+ip·x. (1.239)
176. Using Dirac equation, show that
(p/−m)u(p) = 0, (1.240a)ū(p)(p/−m) = 0, (1.240b)(p/+m)v(p) = 0, (1.240c)
v̄(p)(p/+m) = 0. (1.240d)
177. Using the fact that −p→ p in the replacement v(p) = u(−p), explain why the spin-up positive-energy positron state is expressed in terms of spin-down negative-energy electron state.
178. Formal way to prove this is involving charge-conjuagtion operation with the transformationmatrix
iγ2 =
(0 iσ2
−iσ2 0
)=
0 0 0 10 0 −1 00 −1 0 01 0 0 0
, σ2 = ( 0 −ii 0)
(1.241)
to the spin-half Lagrangian.
179. Replacing the derivative in Dirac equation by the covariant derivative Dµ = ∂µ + iqAµ, showthat
(i∂/− qA/−m)ψ = 0. (1.242)
Taking complex conjugate, show that
(−i∂/∗ − qA/∗ −m)ψ∗ = 0→ [−γµ∗ (i∂µ + qAµ)−m]ψ∗ = 0, (1.243)
where A∗µ = Aµ.
180. Using the fact
γ0 =
(1 00 −1
), γ =
(0 σ−σ 0
)(1.244)
and
σ1 =
(0 11 0
), σ2 =
(0 −ii 0
), σ3 =
(1 00 −1
), (1.245)
show that
−γµ∗ ={−γµ if µ 6= 2,+γµ if µ = 2.
(1.246)
181. Show for
U = iγ2, (1.247)
27
-
that
U2 = 1→ U−1 = U † = U, (1.248a)U−1γµU = −γµ → Uγµ = −γµU if µ 6= 2, (1.248b)U−1γµU = γµ → Uγµ = γµU if µ = 2, (1.248c)
Therefore,iγ2 (−γµ∗) = γµ
(iγ2). (1.249)
182. Show that
iγ2 [−γµ∗ (i∂µ + qAµ)−m]ψ∗ = 0→ [γµ (i∂µ + qAµ)−m](iγ2ψ∗
)= 0. (1.250)
Therefore, iγ2ψ∗ is the wavefunction for the antiparticle.
1.13 Fermion propagator
183. Show that the Dirac(iγµ∂µ −m)ψ = J (1.251)
Be careful with the sign in front of the source term on the right-hand side; H = i∂0.
184. Show that the propagator satisfies the equation
(iγµ∂µ −m)SF (x) = Iδ(x) (1.252)
where I is the 4× 4 identity matix in spinor space.
185. Replacing SF (x) = (iγµ∂µ +m) f(x), show that f(x) satisfies the equation(
∂2 +m2)f(x) = −δ(x) (1.253)
186. Show that the solution to the above equation is
SF (x) = (iγµ∂µ +m) ∆F (x) (1.254)
187. Show that
SF (x) =
∫d4p
(2π)4(/p+m) e−ip·x
p2 −m2 + i�(1.255)
188. Show that the Diracψ̄(−iγµ
←∂ µ −m
)= J (1.256)
where A←∂ µ≡ ∂µA. Be careful with the sign in front of the source term on the right-hand side;
H = i∂0.
189. Show that the propagator satisfies the equation
S′F (x)
(−iγµ
←∂µ −m
)= Iδ(x) (1.257)
where S′F is the propagator for the ψ̄ field. We will show that S′F (x) 6= SF (x)
28
-
190. Replacing S′F (x) = g(x)(−iγµ
←∂ µ +m
), show that g(x) satisfies the equation
(∂2 +m2
)g(x) = −δ(x) (1.258)
191. Show that the solution to the above equation is
S′F (x) = ∆F (x)(−iγµ
←∂ µ +m
)(1.259)
192. Show that
S′F (x) =
∫d4p
(2π)4(−/p+m) e−ip·x
p2 −m2 + i�(1.260)
=
∫d4p
(2π)4(/p+m) eip·x
p2 −m2 + i�= SF (−x) 6= SF (x) (1.261)
(e−ip·x 6= eip·x)
193. Show that SF is not an even function
SF (−x) 6= SF (x), SF (−p) 6= SF (p) (1.262)
194. Show that
ψ(x) =
∫d4ySF (x− y)J(y) =
∫d4ySF (x− y)[qγµAµ(y)]ψ(y), (1.263)
ψ̄(x) =
∫d4yJ(y)SF (y − x) =
∫d4yψ̄(y)[qγµAµ(y)]SF (y − x) (1.264)
195. Show that the positive-energy component propagates as
SF (x) = θ(x0)Sret.(x) + θ(−x0)Sadv.(x) (1.265)
Sret.(x) = −i∫
d3p
(2π)3/p+m
2Ee−ip·x (1.266)
Sadv.(x) = −i∫
d3p
(2π)3−/p+m
2Eeip·x (1.267)
where p0 = E =√m2 + p2 > 0 From now on we neglect the normalization factor N = 1/
√V
which does not affect the invariant measurables like cross section.
196. Transition amplitude M Once we know the transition amplitudeM in momentum space, wecan calculate the cross section of a process.
197. For a scalar-exchange process
M =[−iĴ(−p)
] ip2 −m2 + i�
[−iĴ(p)
](1.268)
29
-
198. For a photon-exchange process
M =[−iĴµ(−p)
] i [−gµν + (1− α)pµpνp2]
p2 + i�
[−iĴν(p)
](1.269)
199. For a massive-vector-exchange process
M =[−iĴµ
] i [−gµν + pµpνm2]
p2 −m2 + i�
[−iĴν
](1.270)
200. For a fermion-exchange process
M =[−iĴ
] i (/p+m)p2 −m2 + i�
[−iĴ
](1.271)
30
-
Chapter 2
QCD Lagrangian
2.1 QED Summary
1. In the previous chapter we wrote the QED interaction Lagrangian, derived Feynman rules, andlearned how to write the amplitude. QED Lagrangian is given by
L = ψ (iγµDµ −m)ψ −1
4FµνF
µν , (2.1)
where the covariant derivative is defined by
Dµ = ∂µ + iqAµ (2.2)
with q = Qe, e =√
4πα > 0 and Q = −1 for the electron.
2. Wavefunctions for incoming positive-energy electron and positron with momentum p are
e− : u(p)e−ip·x, e+ : v̄(p)e−ip·x. (2.3)
3. Wavefunctions for outgoing positive-energy electron and positron with momentum p are
e− : ū(p)e+ip·x, e+ : v(p)e+ip·x. (2.4)
4. Wavefunction for incoming and outgoing photons with momentum k are
in : �µ(k)e−ik·x, out : �µ∗(k)e+ik·x. (2.5)
5. Propagator for the electron with momentum p is
iSF (p) =i
p/−m+ i�. (2.6)
6. Propagator for the photon has various forms depending on the gauge, which does not changethe observables. In the Feynman gauge, the propagator for a photon with momentum k is
iDµνF (k) =−igµν
k2 + i�. (2.7)
31
-
7. The vertex factor can be read from the interaction Lagrangian iLint as
ēAµe = +iqγµ, q = −e. (2.8)
8. Show that once the coupling q is defined by the covariant derivative, the coupling is sameinclusing sign for both electron and positron.
9. We can re-express the field strength tensor in the form
Fµν = ∂µAν − ∂νAµ = 1iq
[Dµ,Dν ] , Dµ = ∂µ + iqAµ. (2.9)
10. Check the gauge invariance of the Lagrangian.
L = ψ (iγµDµ −m)ψ −1
4(iq)2Tr ([Dµ,Dν ] [Dµ,Dν ]) , (2.10)
under the gauge transformation
ψ → Uψ, Dµ → UDµU †. (2.11)
Note that the trace is for the 1× 1 matrix U .
2.2 QCD Lagrangian
11. It is known that there are three (Nc = 3) color states for a quark, which is a spin-half particle.The color is independent of spin and momentum.
12. We can extend QED to treat this new degree of freedom by introducing gauge fields mediatingthe color force between any two quarks. This can be done by declaring the spinor field ψ hasthe wavefunction of the form
ψquark = ψquark(spin)× ψquark(color). (2.12)
13. We can introduce the gauge transform for the quark just like that for the electron as
ψ → Uψ, Dµ → UDµU †. (2.13)
Note that the matrix U is now a 3× 3 matrix and acts only on the color wavefunction.
Uψquark = ψquark(spin)× Uψquark(color). (2.14)
14. We know that U must be unitary. Show that the matrix U can be parametrized by
U = e−i∑N2c−1a=1 T
aαa , (2.15)
where αa is real and T a’s are the SU(Nc) generators.
15. Show that T a is traceless and Hermitian.
32
-
16. Show that the number of generators for the SU(Nc) is N2c − 1.
17. The covariant derivative can be generalized to the SU(Nc) as
Dµ = ∂µ + igsAµ, (2.16)
where Aµ is the matrix-valued gluon field
Aµ = AµaTa. (2.17)
Note that gs is the strong coupling. Therefore, gluons can have N2c −1 = 8 different color states.
18. We can imagine the QCD Lagrangian should be of the form
L = ψ (iγµDµ −m)ψ −1
4Gµνa G
aµν , (2.18)
because there are 8 gluons. Note that Gµνa is the field strength tensor for the gluon with colora, which is a QCD analogy of the photon field strength tensor.
19. Let us check the gauge invarince of the Lagrangian. We can use the fact
Tr(T aT b
)=
1
2δab, (2.19)
to derive
−14Gµνa G
aµν = −
1
2Tr (GµνGµν) . (2.20)
Again, Gµν is a matrixGµν = Gµνa T
a. (2.21)
20. Show that the covariant derivative transforms as
Dµ → UDµU †, DµDν → UDµDνU †, (2.22)
under the gauge transformation
ψ → Uψ, Aµ → UAµU † − 1igs
(∂µU)U †. (2.23)
21. Show that the QCD Lagrangian
L = ψ (iγµDµ −m)ψ −1
2Tr (GµνGµν) (2.24)
is invariant under the gauge transformation. We have constructed the QCD Lagrangian, whichis gauge invariant and Lorentz invariant.
22. In the previous section, we have written the QED Lagrangian as the following.
L = ψ (iγµDµ −m)ψ −1
4Gµνa G
aµν . (2.25)
Show that
Gµν =1
igs[Dµ,Dν ] = ∂µAν − ∂νAµ + igs [Aµ, Aν ] . (2.26a)
33
-
23. Show thatGµνa = 2Tr (G
µνT a) = ∂µAνa − ∂νAµa + 2igsAµbA
νcTr
(T a[T b, T c]
). (2.27)
24. Using the identity [T b, T c] = ifabcT a, show that
Gµνa = ∂µAνa − ∂νAµa − gsfabcA
µbA
νc . (2.28)
25. Show that there exists three-gluon coupling in QCD.
26. Show that there exists four-gluon coupling in QCD.
2.3 Gauge Fixing in QED
27. We learned that we need gauge-fixing term such as
Lgauge-fixing = −1
2α(∂µA
µ)2 (2.29)
in the Lorentz gauge in order to derive the photon propagator.
28. Show in QED that the gauge transform changes the photon field as
Aµ + ∂µχ, (2.30)
if we chooseU = e−iqχ. (2.31)
29. Show that the gauge condition ∂µAµ = 0 transforms as
∂µAµ + ∂µ∂
µχ = 0. (2.32)
There appears unpleasant term ∂µ∂µχ.
30. Show that∂µ∂
µχ = 0, (2.33)
if we want to keep the gauge condition ∂µAµ = 0. Under the transform, gauge field is shifted to
another gauge field which still satisfies the condition ∂µAµ = 0.
31. Once we introduce a gauge-fixing term, our gauge transform cannot be completely general. Thetransform must keep the gauge condition in order for the Lagrangian to be gauge invariant. Thegauge invariance is valid only in the group of gauges satisfying the same gauge condition.
32. Show that∂µ∂
µχ = 0, (2.34)
is the equation of motion for the Lagrangian
L = ∂µχ†∂µχ. (2.35)
We can include this term into the Lagrangian in order to correct the gauge invariance of ∂µAµ.
33. Show that inclusion of this scalar field does not make any physical contribution in reality becauseit does not have any interaction term with any other fields.
34
-
2.4 Gauge-Fixing and Ghost Terms in QCD
34. We can introduce the same kind of gauge-fixing term like that of the Lorentz gauge in QED. Wecan try the QCD analogy of this
Lgauge-fixing = −1
2α(∂µA
µa)
2 . (2.36)
Note that there are N2c − 1 conditions ∂µAµa = 0 for a =1, 2, · · · , N2c − 1.
35. We learned that in QCD the gauge transform changes the gluon field as
Aµ → UAµU † − 1igs
(∂µU)U †, (2.37)
where Aµ = AµaT a. Let us find how each gluon field Aµa transforms in QCD.
36. It is convenient to choose the parametrization as
U = e−igsαaTa . (2.38)
Note that αa should be real to preserve U unitary. Expanding the matrix U in powers of α uptocorrections of order α2, show that
Aµ → (1− igsαcT c)AµxT x (1 + igsαcT c) + ∂µαaT a (2.39a)= (Aµa + ∂
µαa)Ta + igsA
µx[T
x, T c]αc (2.39b)
= (Aµa + ∂µαa)T
a + igsAµx (if
axc)αcTa. (2.39c)
Therefore, Aµa transforms as
Aµa → Aµa + [∂µδac + igsAµx (ifaxc)]αc. (2.40)
37. Let us introduce a matrix in the adjoint representation
(tx)ac ≡ ifaxc. (2.41)
Show thatAµa → Aµa + [∂µδac + igsAµx (tx)ac]αc. = A
µa + D̃µacαc, (2.42)
where the covariant derivative D̃ac in the adjoint representation has the same form
D̃µac ≡ ∂µδac − gsAµxfaxc = ∂µδac + igsõac, õac ≡ Aµx (tx)ac = Aµx (if
axc) . (2.43)
as that in the fundamental represenation.
38. The problem is resolved if we include the ghost term to the Lagrangian. One can find detaileddiscussion in most field theory text book such as Peskin.
Lghost = (∂µη̄a) D̃µacηc = (∂µη̄a) (∂µδac − gsfaxcAµx) ηc. (2.44)
39. Ghost η is a complex scalar field. However, it behaves like a fermion in statistical sense. We willfind later.
35
-
2.5 SU(Nc) algebra summary
SU(Nc) Generator Ta, a = 1, · · · , N2c − 1
[T a, T b] = ifabcT c (2.45)
(tb)ac = ifabc (2.46)
Fij = FaT aij ← F a = 2 Tr(FT a), Tr(T a, T b) =
1
2δab (2.47)
faxyf bxy = Ncδab
= −(ta)xy(tb)xy = (ta)xy(tb)yx = Tr(tatb) (2.48)
2.6 QCD Feynman rules
40. we are ready to derive QCD Feynman rules from the QCD Lagrangian.
L = ψ (iγµDµ −m)ψ −1
4
(∂µAνa − ∂νAµa − gsfabcA
µbA
νc
)(∂µA
aν − ∂νAaµ − gsfapqApµAqν
). (2.49)
where the covariant derivatives are defined by
Dµij = δij∂µ + igsA
µaT
aij (2.50)
Dµac = δab∂µ + igsAµb (t
b)ac = δab∂µ − gsGµb f
abc ← (tb)ac = ifabc (2.51)
where indices i, j, k, · · · and a, b, c, · · · denote color indices for quark and gluon, respectively.
41. Let us recall the followings. Field strength tensor from fundamental representation
Gµν = − igs
[Dµ,Dν ] = − igs
[∂µ + igsAµaTa, ∂
ν + igsAνbTb]
= ∂µAνaTa − ∂νAµaTa + igsAµxAνy [Tx, Ty]= ∂µAνaT
a − ∂νAµaT a − gsAµxAνyfxyaT a ← [T a, T b] = ifabcT c (2.52)Gµνa = 2Tr(G
µνT a) = ∂µAνa − ∂νAµa − gsfabcAµbA
νc (2.53)
(tb)ac = ifabc (2.54)
42. Derivative
∂µ → −ipµ(incoming line), ∂µ → +ipµ(outgoing line) (2.55)
43. Vertex
+i L (2.56)
44. Wavefunctions for incoming positive-energy quark and antiquark with momentum p are
q : u(p)e−ip·x, q̄ : v̄(p)e−ip·x, (2.57)
where we have suppressed the color wavefunction. The color wavefunction is of the form
ci, c†icj = δij , i, i = 1, 2, 3. (2.58)
36
-
45. Wavefunctions for outgoing positive-energy quark and antiquark with momentum p are
q : ū(p)e+ip·x, q̄ : v(p)e+ip·x. (2.59)
46. Wavefunction for incoming and outgoing gluons with momentum k are
in : �µ(k)e−ik·x, out : �µ∗(k)e+ik·x. (2.60)
The color wavefunction for the gluon with color index a is
ca, c†acb = δab, i, i = 1, 2, ·, 8. (2.61)
47. Propagator for the quark with momentum p with initial and final color indices i and j is
iSF (p)× δji =iδji
p/−m+ i�. (2.62)
The factor shows the color is preserved δji.
48. Propagator for the photon has various forms depending on the gauge, which does not changethe observables. In the Feynman gauge, the propagator for a photon with momentum k is
iDµνF (k)× δba =−igµνδbak2 + i�
. (2.63)
49. The q̄jAµaqi vertex factor can be read from the interaction Lagrangian iLint as
q̄jAµaqi : −igsT ajiγµ. (2.64)
50. Coupling gs is universal for any quark.
2.6.1 Gluon Vertices
51. From the QCD Lagrangian
L = −14
(∂µAνa − ∂νAµa)2 −1
2α(∂ ·Aa)2
+1
2gsf
abc(∂µAνa − ∂νAµa)AbνAcν −1
4g2sf
xabfxcdAµaAνbA
cµA
dν , (2.65)
derive the following Feynman rules.
52. gluon propagrator
g : − igµν − λ−1λ p
µpν/p2
p2 + i�(2.66)
53. Three-gluon vertex gµ1c1 (p1 : in)− gµ2c2 (p2 : in)− g
µ3c3 (p3 : in) is
ggg : − gsf c1c2c3 ((p1 − p2)µ3gµ1µ2 + (p2 − p3)µ1gµ2µ3 + (p3 − p1)µ2gµ3µ1) . (2.67)
37
-
54. Solution:
iLggg = i[−1
4GaµνG
µνa
]ggg
= igfabcAµbAνc∂µA
aν (2.68)
= g∑
perm.{1,2,3}
fabc∫d4p1d
4p2d4p3e
−i(p1+p2+p3)·x(p1 ·Aa2)(Ab1 ·Ac3) (2.69)
= −g∑
perm.{1,2,3}
fabc∫d4p1d
4p2d4p3e
−i(p1+p2+p3)·x
×[−(p1 ·Aa2)(Ab1 ·Ac3) + (p3 ·Aa2)(Ab1 ·Ac3) · · ·
]=
∫d4p1d
4p2d4p3e
−i(p1+p2+p3)·xAaµ1(p1)Abµ2(p2)A
cµ3(p3)
×(−gfabc) · [gµ1µ2(p1 − p2)µ3 + gµ2µ3(p2 − p3)µ1 + gµ3µ1(p3 − p1)µ2 ] . (2.70)
(abc) = (c1c2c3) If we choose the momentum direction into the vertex(annihilation at x), mo-mentum dependence is e−ik·x style and derivative can be replaced by −ik in momentum space.
55. Following the above way, derive the four-gluon vertex gµ1c1 (p1 : in) − gµ2c2 (p2 : in) − g
µ3c3 (p3 :
in)− gµ4c4 (p4 : in) is
gggg : − ig2s[
+f c1c2xf c3c4x (gµ1µ3gµ2µ4 − gµ1µ4gµ2µ3)−f c1c3xf c2c4x (gµ1µ4gµ2µ3 − gµ1µ2gµ3µ4)+f c1c4xf c2c3x (gµ1µ2gµ3µ4 − gµ1µ3gµ2µ4)
]. (2.71)
56. Show the Bose symmetry of three- and four-gluon vertices explicitly.
2.6.2 Ghost Vertices
57. From the ghost term
L = −δacηa∂2ηc + gsfabcηa∂µAµb ηc, (2.72)
derive the following rules
58. ghost(p) propagator
gh :i
p2 + i�(2.73)
59. ghost(pfout, cf )− g(µ, cg)− ghost(piin, color = ci)
ghf − g − ghi : + pµfgsfcf cgci (2.74)
38
-
Chapter 3
SU(N)
In this chapter, we review the properties of the special unitary group, SU(N), which is useful incalculating color factors in various QCD processes.
3.1 Generators and structure constants
1. Consider a transform N ×N matrix U which transforms a matrix O and a vector ψ in complexfield as
ψ → ψ′ = Uψ and O → O′ = UOU−1. (3.1)
Show that ψ†ψ is invariant under this transformation
ψ†Oψ → ψ†U †UOU−1Uψ = ψ†Oψ, (3.2)
if the transformation operator is unitary:
U−1 = U †. (3.3)
When the transform is infinitesimal, U may be expressed as
U = e−i�aTa = 1− i�aTa, (3.4)
where �a’s are the infinitesimal real parameters and Ta’s are the generators of the transformation.
2. Show that iT a must be antihermitian.
3. Show that T a must be hermitian.
4. Show that T a must be traceless. Since U is unitary, Ta’s are Hermitian. If we restrict detU = +1as the case of the identity transformation, the Ta matrices are restricted to be traceless as
det U = �i1i2,...,iNU1i1U2i2 ... UNiN= �i1i2,...,iN (δ1i1 − i�aT
1i1a )(δ2i2 − i�aT 2i2a ) ... (δNiN − i�aT
1iNa )
= �12,...,N − i�a(�i1i2,...,iN (T
1i1a δ1i1 ... δNiN + ...+ δ1i1 ... δN−1iN−1T
1i1a )
)= 1− i�a
(�i12,...,NT
1i1a + ...�12,...,iNT
NiNa
)= 1− i�aTrTa → TrTa = 0 ← detU = +1. (3.5)
39
-
5. Show that there are N2c −1 free real parameters for �a. Consider how many Ta’s are independent.Since a transformation matrix U is an N×N complex matrix, there are 2N2 parameters at first.Hermitian constraint discards N2 parameters and traceless condition does one more parameterso that we now have N2 − 1 independent Ta’s.
6. If N2 = 2, the problem is the same as that for a spin-1/2 particle. Show that the generators forthe SU(2) are Pauli sigma matrices. Are there N2c − 1 = 3 generators?
7. Consider a commutator −i [Ta, Tb]. It is a traceless and hermitian operator. Therefore it can beexpressed as a linear combination of the generators as
[Ta, Tb] = ifabcTc, (3.6)
where fabc is the ab anti-symmetric evidently from the definition using commutator.
8. Show that fabc is totally anti-symmetric.
fabc = f bca = f cab
= −f bac = −f cba = −facb = 13!
(fabc + f bca + f cab − f bac − f cba − facb
). (3.7a)
You can prove it if you use the relations
Tr[AB] = Tr[BA]→ Tr[ABC] = Tr[BCA] = Tr[CAB], (3.8)
which are valid for any matrices.
9. Since the trace of any two generator product is symmetric under the exchange of the colorindices,
Tr[T aT b] = Tr[T bT a] =1
2
(Tr[T aT b] + Tr[T bT a]
), (3.9)
Tr[T aT b] is proportional to δab and the normalization can be chosen freely. Conventionally λ isset to 1/2.
Tr [TaTb] =1
2δab. (3.10)
10. We can infer that the antisymmetric term in TaTb is proportional to ifabcTc from the commutationrelation. Therefore we can express the product of any two generators as
TaTb =1
2[(Afabc +Bdabc)Tc + CδabI] , (3.11)
where dabc is ab-symmetric and A, B and C are to be determined as follows. Directly from thecommutation relation, we can find A = i and from the normalization condition, Tr[TaTb] = δab/2,we can find C = 1/N . Setting B = 1 also, we get
TaTb =1
2
[(ifabc + dabc)Tc +
1
NδabI
],
{Ta, Tb} = dabcTc +1
NδabI. (3.12)
40
-
11. Then we find that dabc is fully symmetric as
Tr [{Ta, Tb}Tc] = dabdTr [TdTc] =1
2dabc ← ab − symmetric
= Tr [TaTbTc + TbTaTc] = Tr [TbTcTa + TbTaTc]
= Tr [Tb {Ta, Tc}] ← ac − symmetric. (3.13)
12. And these are the useful relations
Tr[T aT b] =1
2δab, (3.14)
Tr[T aT bT c] =1
4[dabc + ifabc], (3.15)
dabc = 2Tr[{T a, T b}T c], (3.16)fabc = −2iTr[[T a, T b]T c], (3.17)
Tr[T aT bT cT d] =1
4
[1
Nδabδcd +
1
2(dabe + ifabe)(dcde + if cde)
]. (3.18)
3.2 Derivation of completeness relation
Any N ×N traceless and hermitian matrix A can be expressed in the linear combination of thegenerators as
A =N2−1∑a=1
αaT a, (3.19)
where αa’s are real. There are N2 degrees of freedom in arbitrary N × N hermitian matricesand there are N2 − 1 generators, T a’s. Therefore, we need one more N × N matrix to form abasis of the hermitian matrix other than the N2 − 1 generators, T a’s. Since identity matrix isan hermitian and it is independent of the generators, we can form a basis by adding the identitymatrix.
13. Show that any N × N Hermitian matrix is expressed in a linear combination of the identitymatrix and SU(N) generators.
H = α0I +N2−1∑a=1
αaT a, (3.20)
where αi for i =0, 1, · · · , N2 − 1 are all real. Then we can obtain useful relation from thiscompleteness condition. By choosing the normalization
Tr [TaTb] =1
2δab, (3.21)
we obtain the explicit values of the coefficients as
TrH = α0N and Tr[T aH] = 12αa. (3.22)
41
-
14. Rewriting the hermitian matrix H
H = 1N
Tr[H] I + 2N2−1∑a=1
Tr[T aH]T a, (3.23)
in matrix representation, show that
Hµν =1
NHααδµν + 2
N2−1∑a=1
T aµνTaαβHβα,
Hβα(δµβδνα) = Hβα
1Nδαβδµν + 2
N2−1∑a=1
T aµνTaαβ
. (3.24)15. Using the fact that Hµν is an arbitrary complex number for any µ > ν and an arbitrary real
number for any µ = ν, prove the completeness relation
N2−1∑a=1
T aµνTaαβ =
1
2
(δµβδνα −
1
Nδµνδαβ
). (3.25)
With this relation, we can calculate any color factor involving SU(N) gauge theory.
3.3 Useful trace formulas
We can derive various trace formulas and relations among the structure constants which are veryuseful in practical calculations concerning perturbative QCD. In this section, we derive thesepractically useful relations in detail, by using the results shown in previous sections.
From now on, we use summation convention, where any two repeated indices are assumed to besummed over color indices.
16. By using the completeness relation, we can show that the sum of squared generators is propor-tional to the identity matrix asN2−1∑
a=1
T aT a
µν
=
N2−1∑a=1
T aµαTaαν =
1
2
(δµνδαα −
1
Nδµαδαν
)
=1
2
(Nδµν −
1
Nδµν
),
→N2−1∑a=1
T aT a =N2 − 1
2NI. (3.26)
For SU(N = 3),
CF =N2 − 1
2N=
4
3. (3.27)
The color factor appears in the quark wavefunction renormalization factor.
42
-
17. When we calcualte gluon-loop corrections to ferimion-gluon-fermion vertices, we need to evaluatethe matrix such as ∑
a
T aT bT a. (3.28)
Using the completeness relation to re-order the matrix product and using the formula∑a
T aT a =N2 − 1
2NI, (3.29)
show that
T aT aT b =N2 − 1
2NT b, (3.30)
T aT bT a = − 12N
T b, (3.31)
T aT aT bT b =(N2 − 1)2
4N2I, (3.32)
T aT bT aT b = −N2 − 1
4N2I, (3.33)
T aT bT bT a =(N2 − 1)2
4N2I. (3.34)
18. We can derive the following trace formulas using the same method used above. Here are theSU(N) trace formulas up to 4 pairs of indices
Tr[T aT b] =1
2δab, (3.35)
Tr[T aT aT b] = 0, (3.36)
Tr[T aT b]Tr[T aT c] =1
4δbc, (3.37)
Tr[T aT aT bT c] =N2 − 1
4Nδbc, (3.38)
Tr[T aT bT aT c] = − 14N
δbc, (3.39)
Tr[T aT bT c]Tr[T aT bT d] = − 14N
δcd, (3.40)
Tr[T aT bT cT aT bT d] =N2 + 1
8N2δcd, (3.41)
Tr[T aT bT c]Tr[T bT aT d] =N2 − 2
8Nδcd, (3.42)
Tr[T aT bT cT bT aT d] =1
8N2δcd, (3.43)
Tr[T aT aT bT bT cT d] =(N2 − 1)2
8N2δcd, (3.44)
Tr[T aT aT bT cT bT d] = −N2 − 1
8N2δcd, (3.45)
Tr[T aT bT aT bT cT d] = −N2 − 1
8N2δcd. (3.46)
43
-
3.4 Adjoint representation
In this section we investigate the properties of the structure constant fabc’s and symmetric dabc’s.And we will look into the adjoint representation which is made up of the structure constant fabc
itself.
Compare the above identity with the one driven in previous section,
T aT b =1
2
[1
Nδab I + (dabc + ifabc)T c
]. (3.47)
19. If we multiply δab and sum over color indices, then we get the properties of the structure constantand symmetric dabc as
faab = 0 and daab = 0. (3.48)
20. If we use Eq.(3.47) two times we get
T aT bT c =1
4N[dabc + ifabc] I
+1
2
[1
Nδabδce +
1
2(dabd + ifabd)(ddce + ifdce)
]T e. (3.49)
Let us consider triple product T aT bT c. By using Eq. (3.26) we get
T aT aT b =N2 − 1
2NT b and T aT bT a = − 1
2NT b
−→ T a[T a, T b] = N2T b and T a{T a, T b} = N
2 − 22N
T b. (3.50)
Comparing the result by using the Eqs.(3.6), (3.12)
T a[T a, T b] = T aifabcT c =i
2fabc[T a, T c] =
i
2fabcifaceT e =
1
2facbfaceT e,
T a{T a, T b} = T a(dabcT c +
1
NδabI
)=
1
2dabc{T a, T c}+ 1
NT b
=1
2dabcdaceT e +
1
NT b, (3.51)
show that
fabcfabd = Nδcd (3.52a)
fabcdabd = 0 (3.52b)
dabcdabd =N2 − 4N
δcd. (3.52c)
21. With the identities and Eqs. (3.16) and (3.17) we obtain other relations
dabcfabd = 0,
dabcdabc =(N2 − 1)(N2 − 4)
N,
fabcfabc = N(N2 − 1). (3.53)
44
-
22. The structure constants fabc themselves define adjoint representation
F bac ≡ ifabc. (3.54)
23. Show that F b’s are traceless, hermitian, and anti-symmetric (N2 − 1)× (N2 − 1) matrices.
24. By using the Jacobi identity, show that
[[T a, T b], T c] + [[T b, T c], T a] + [[T c, T a], T b] = 0,
ifabd[T d, T c] + if bcd[T d, T a] + if cad[T d, T b] = 0,
ifabdifdceT e + if bcdifdaeT e + if cadifdbeT e = 0,
(−ifabd)(−ifdce) + (−if bcd)(−ifdae) + (−if cad)(−ifdbe) = 0,(−ifabd)(−ifdce)− (−ifadc)(−ifebd) = ifaed(−ifdbc),
(F aF e)bc − (F eF a)bc = ifaedF dbc→ [F a, F b] = ifabcF c. (3.55)
25. Defining the fully symmetric dabc likewise(Dabc = dabc), we get
[{T a, T b}, T c] + [{T b, T c}, T a] + [{T c, T a}, T b] = 0,dabd[T d, T c] + dbcd[T d, T a] + dcad[T d, T b] = 0,
dabdifdceT e + dbcdifdaeT e + dcadifdbeT e = 0,
dabdifedc − dadcifebd = −dbcdifdae,−(DaF e)bc + (F eDa)bc = ifeadDdbc,
→ [F a, Db] = ifabcDc. (3.56)
26. Formulas related to the structure constants can be rewritten in terms of adjoint representationas
faab = 0 → TrF b = 0, (3.57)daab = 0 → TrDb = 0, (3.58)
fabcfabd = (−if cab)(−ifdba) = Tr[F cF d] = Nδcd, (3.59)
fabcfabd = (−ifacb)(−ifabd) =N2−1∑a=1
(F aF a)cd = N Icd, (3.60)
dabcfabd = 0 → Tr[DcF d] = 0, (3.61)
dabcfabd = 0 →N2−1∑a=1
(DaF a)cd = 0cd, (3.62)
dabcdabd = dcabddba = Tr[DcDd] =N2 − 4N
δcd, (3.63)
dabcdabd = dacbdabd =
N2−1∑a=1
(DaDa)cd =N2 − 4N
Icd. (3.64)
45
-
Now we summarize the results concerning the two representations. Regardless of the two repre-sentations
T (R) δab = Tr [TaTb] and C2(R) I =N2−1∑a=1
T (R)a T(R)a , (3.65)
where I is the identity matrix in the representation R and the label R is given as
T (F )a = Ta and T(A)a = Fa. (3.66)
We can set the normalization of the generators by setting the value of T (R). Standard normal-izations are given as
T (R) =1
2and T (R) = N. (3.67)
Then the values of C2(R) are fixed as
C2(R) =N2 − 1
2Nand C2(R) = N. (3.68)
3.5 SU(3) Clebsch-Gordan coefficients
27. As the meson formed from qq into singlet and octet in flavor SU(3), the color state formed byQ and Q with color i and j form a color singlet and a color octet states as
3⊗ 3̄ = 1⊕ 8. (3.69)
Obviously the singlet state is the identity matrix and octet states are the eight generators ofSU(3) in the fundamental representation up to normalization as
〈3i; 3j|1〉 = N1δij and 〈3i; 3j|8a〉 = N8T aij . (3.70)
The octet coefficient is proportional to the physical vertex ; gluon with color a goes to a color iquark and a color j antiquark. One should be cautious not to confuse the order of indices i andj in above equations.
Normalizing the nine states as
〈1|1〉 =∑ij
〈1|3i; 3j〉〈3i; 3j|1〉
= |N1|2∑ij
δjiδij
= |N1|2Tr[II] = 3|N1|2 = 1, (3.71)〈8a|8a〉 =
∑ij
〈1|3i; 3j〉〈3i; 3j|1〉 (not sum over index a)
= |N8|2∑ij
T a∗ij Taij
= |N8|2∑ij
T ajiTaij
= |N8|2Tr[T aT a] =1
2|N8|2 = 1. (3.72)
46
-
If we perform the same derivation for the case of SU(N), then we get
〈3i; 3j|1〉 = 1√Nδij and 〈3i; 3j|8a〉 =
√2T aij , (3.73)
provided that the generators of the fundamental representation are normalized as
Tr[T aT b] =1
2δab. (3.74)
28. To find the color factor for singlet and octet in this case, we use this kind of method: Since anyproduct of color matrices are expressed as
Tn ≡ T a1T a2T a3 · · · ·T an = A(a1, a2, . . . , an) +Ba(a1, a2, . . . , an)T a (3.75)
A(a1, a2, . . . , an) =1
NTr(Tn) (3.76)
Ba(a1, a2, . . . , an) = 2Tr(TaTn) (3.77)
Then the total color factor is
CT ≡ TrTnT †n = A(a1, a2, . . . , an)A∗(a1, a2, . . . , an)Tr(I)+ Ba(a1, a2, . . . , an)B
∗b(a1, a2, . . . , an)Tr(TaT b)
= A(a1, a2, . . . , an)A∗(a1, a2, . . . , an)N
+ Ba(a1, a2, . . . , an)B∗b(a1, a2, . . . , an)
1
2δab
= NA(a1, a2, . . . , an)A∗(a1, a2, . . . , an)
+1
2Ba(a1, a2, . . . , an)B
∗a(a1, a2, . . . , an) (3.78)
And the singlet and octet factors are
C1 = A(a1, a2, . . . , an)A(a1, a2, . . . , an)
=1
N2Tr(Tn)Tr(Tn) (3.79)
C8δab = Ba(a1, a2, . . . , an)B
b(a1, a2, . . . , an)
=δab
N2 − 1Ba(a1, a2, . . . , an)B
a(a1, a2, . . . , an) (3.80)
C8 =1
N2 − 1Ba(a1, a2, . . . , an)B
a(a1, a2, . . . , an)
=4
N2 − 1Tr(T aTn)Tr(T
aTn) (3.81)
Therefore there is a relation among the total color factor, singlet and octet color factor
CT = NC1 +1
2(N2 − 1)C8 (3.82)
47
-
3.6 Examples
29. In calculations of QCD amplitudes such as gg → gg, we have to calculate color factors such as
fabcfaxyfdbxfdcz, (3.83)
which is very involved.
fabcfaxyfdbxfdcz = +1
2N2δyz (3.84)
dabcfaxyfdbxfdcz = 0 (3.85)
dabcdaxyfdbxfdcz = +1
2(N2 − 4)δyz (3.86)
dabcfaxyddbxfdcz = −12
(N2 − 4)δyz (3.87)
dabcdaxyddbxfdcz = 0 (3.88)
dabcdaxyddbxddcz = +1
2N2(N2 − 4)(N2 − 12)δyz (3.89)
where the last formula is not derived directly from the relations given above, instead, with
44 × Tr(T aT bT c)Tr(T aT xT y)Tr(T dT bT x)Tr(T dT cT z)= (dabc + ifabc)(daxy + ifaxy)(ddbx + ifdbx)(ddcz + ifdcz) (3.90)
and the remaining formulas. And the relation is invariant under cyclic rotations such as
AabcBaxyCdbxDdcz = BabcCaxyDdbxAdcz
= CabcDaxyAdbxBdcz
= DabcAaxyBdbxCdcz (3.91)
30. The completeness relation is the most powerful tool in calculating color factors. You are ad-vised to transform any color factors into the fundamental representation first. Then using thecompleteness relation to reorder the color matrices so that you can express the color factor as alinear combination of products of color factors made of
Tr[T aT b] =1
2δab. (3.92)
Then the remaining calculations are products of
δaa = N2 − 1 and/or δii = N. (3.93)
48
-
Chapter 4
Tree-Level Calculation
4.1 Cross section formula Summary
1. Discuss the difference between the two matrices T and M.
S = 1 + iT , (4.1)
2. In the transition matrix T , scalar wavefunction factor
e−ip·x (4.2)
for each particle is included.
3. When we obtain the transition amplitudeM, position variable in T for each vertex is integratedout to give energy-momentum conservation delta function∫
d4xei∑i pi·x = (2π)4δ(4)
(∑i
pi
). (4.3)
4. When we square the matrix T , we treat the space-time volume finite TV .[(2π)4δ(4)
(∑i
pi
)]2= TV × (2π)4δ(4)
(∑i
pi
)(4.4)
5. In the transition matrix T , each external particles have the normalization to get∫ψ†(x)ψ(x)d3x =
2E
V× V = 2E, (4.5)
6. Show that the cross section formula for the following process
A(p1,M1) +B(p2,M2)→ f1(q1,m1) + f2(q2,m2) + · · ·+ fn(qn,mn), (4.6)
49
-
is given by
dσ =
∑M2
FluxdΦn. (4.7a)
Flux = 4√
(p1 + p2)2 −M21M22 , (4.7b)
dΦn = (2π)4δ(4)
(p1 + p2 −
n∑k=1
qk
)Πnk=1
dqk2Ei(2π)3
, (4.7c)
Ek =√m2k + q
2k. (4.7d)
The notation∑
stands for
1
(2JA + 1)(2JB + 1)NA(color)NB(color)
∑spinA
∑colorA
∑spinB
∑colorB
Πnk=1∑spink
∑colork
(4.8)
summation over all the final states and averaging over all the initial states. The states includesall possible degress of freedom such as spin and color.
7. Show that the following average factors
e+e− → f1 + · · ·+ fn =(
1
2
)e+−spin
(1
2
)e−−spin
=1
4, (4.9a)
e−γ → f1 + · · ·+ fn =(
1
2
)e−−spin
(1
2
)γ−spin
=1
4, (4.9b)
γγ → f1 + · · ·+ fn =(
1
2
)γ−spin
(1
2
)γ−spin
=1
4, (4.9c)
qq̄ → f1 + · · ·+ fn =(
1
2
)q−spin
(1
3
)q−color
(1
2
)q̄−spin
(1
3
)q̄−color
=1
36, (4.9d)
qg → f1 + · · ·+ fn =(
1
2
)q−spin
(1
3
)q−color
(1
2
)g−spin
(1
8
)g−color
=1
96, (4.9e)
gg → f1 + · · ·+ fn =(
1
2
)g−spin
(1
8
)g−color
(1
2
)g−spin
(1
8
)g−color
=1
256. (4.9f)
4.2 e+e− → µ+µ−
8. Neglecting masses, show that
σ(e+e− → µ+µ−) = 4πα2
3s. (4.10)
9. Draw the Feynman diagram in the leading order in α.
10. Draw the Feynman diagram in the next-to-leading order in α.
50
-
11. Write the transition amplitude for the process e+(p1)e−(p2) → µ+(q1)µ−(q2) in the light-cone
gauge, n ·A = 0 with n2 = 0.
−iM = [(+ie)v̄(p1)γµu(p2)]× (+i)−gµν + (p1+p2)
µnν+nµ(p1+p2)ν
n·(p1+p2)
(p1 + p2)2 + i�
× [(+ie)ū(q2)γνv(q1)] . (4.11a)
12. Write the same amplitude in the covariant gauge ∂ ·A = 0.
−iM = [(+ie)v̄(p1)γµu(p2)]× (+i)−gµν +
(1α − 1
) (p1+p2)µ(p1+p2)ν(p1+p2)2
(p1 + p2)2 + i�
× [(+ie)ū(q2)γνv(q1)] . (4.12a)
13. Write the same amplitude in the Feynman gauge ∂ · A = 0 with gauge-fixing term − 12α(∂ · A)2
and α = 1.
−iM = [(+ie)v̄(p1)γµu(p2)]× (+i)−gµν
(p1 + p2)2 + i�
× [(+ie)ū(q2)γνv(q1)] . (4.13a)
14. Explain why the three amplitudes must give the same answer.
15. What happens to the gauge-dependent terms proportional to
(p1 + p2)µ(p1 + p2)
ν (4.14a)
(p1 + p2)µnν + nµ(p1 + p2)
ν (4.14b)
in the propagator in the long run?
16. Show that the squared amplitude in the Feynman gauge becomes∑spin
|M|2 = e4
[(p1 + p2)2]2Tr [γµ(/p2 +me)γ
ν(/p1 −me)] Tr [γµ(/q2 +mµ)γν(/q1 −mµ)] . (4.15)
17. Usingγµγν + γνγµ = 2gµν , (4.16)
show that
Tr [/a/b] = 4a · b, (4.17a)Tr [/a/b/c/d] = 4 [(a · b)(c · d)− (a · c)(b · d) + (a · d)(b · c)] . (4.17b)
18. Neglecting me and mµ, show that∑spin
|M|2 = 16e4
[(p1 + p2)2]2[pµ1p
ν2 + p
µ2p
ν1 − gµνp1 · p2] [q
µ1 q
ν2 + q
µ2 q
ν1 − gµνq1 · q2] . (4.18)
51
-
19. Using four-momentum conservation
p1 + p2 = q1 + q2, (4.19)
show that
s ≡ (p1 + p2)2 = (q1 + q2)2, (4.20a)t ≡ (p1 − q1)2 = (p2 − q2)2, (4.20b)u ≡ (p1 − q2)2 = (p2 − q1)2, (4.20c)
s+ t+ u = 2m2e + 2m2µ → 0. (4.20d)
20. Show that
pµ1pν2 [q
µ1 q
ν2 + q
µ2 q
ν1 − gµνq1 · q2] =
1
4(t2 + u2 − s2), (4.21a)
pµ2pν1 [q
µ1 q
ν2 + q
µ2 q
ν1 − gµνq1 · q2] =
1
4(t2 + u2 − s2), (4.21b)
−gµνp1 · p2 [qµ1 qν2 + q
µ2 q
ν1 − gµνq1 · q2] =
s
2
(−s
2− s
2+ 4× s
2
)=s2
2. (4.21c)
21. Using above relations, show that
∑|M|2 = 16e
4
s21
2(t2 + u2)
=8(4π)2α2
s2(t2 + u2)← e2 = 4πα, (4.22a)∑
|M|2 = 14
∑|M|2 = 2(4π)
2α2
s2(t2 + u2). (4.22b)
22. Show that the colliding flux is
Flux = 4√
(p1 + p2)2 = 2s. (4.23)
23. Show that the phase space for the two-body final state is Lorentz invariant.
24. Show that the phase space for the two-body final state in the center-of-momentum frame of thecolliding pair becomes
dΦ2 = (2π)4δ(4)(p1 + p2 − q1 − q2)
d3q12E1(2π)3
d3q22E2(2π)3
(4.24a)
=d3q∗
4E∗1E∗2(2π)
2δ(√s− E∗1 − E∗2), (4.24b)
E∗i =√m2µ + q
∗2 = |q∗|, (4.24c)√s = E∗1 + E2∗, (4.24d)
where q∗ is the momentum q1 measured in the center-of-momentum frame.
52
-
25. Neglecting azimuthal (angle φ) dependence, we can write
d3q∗ = |q∗|2d|q∗|dΩ∗ = 2π|q∗|2d|q∗|d cos θ∗. (4.25)
Using the relation,
δ(f(x)) =∑i
δ(x− xi)|f ′(xi)|
, (4.26)
for a function made of single zeros at x = xi, show that∫d3q∗δ(
√s− E∗1 − E∗2) =
∫|q∗|2d|q∗|dΩ∗δ(
√s− E∗1 − E∗2)
=|q∗|2dΩ∗|q∗|E∗1
+ |q∗|
E∗2
= E∗1E∗2
|q∗|√sdΩ∗. (4.27a)
26. Show finally that
dΦ2 =1
8π× 2|q
∗|√s× dΩ
∗
4π. (4.28)
27. For massless final states, show that
dΦ2 =1
8π× dΩ
∗
4π=d cos θ∗
16π. (4.29)
28. If the amplitude is independent of Ω∗ (rotationally symmetric), show that
dΦ2 =1
8π. (4.30)
29. Combining previous results, show that
σ =
(1
2s
)Flux
∫ 1−1
d cos θ∗
16π
[2(4π)2α2
s2(t2 + u2)
]∑|M|2
. (4.31a)
30. In the center-of-momentum frame, show that
p1 =
√s
2(1, 0, 0, 1), (4.32a)
p2 =
√s
2(1, 0, 0, 1), (4.32b)
q1 =
√s
2(1, sin θ∗ cosφ∗, sin θ∗ sinφ∗, cos θ∗), (4.32c)
q2 = (1,− sin θ∗ cosφ∗,− sin θ∗ sinφ∗, cos θ∗). (4.32d)
Therefore,
t = −2p1 · q1 = −s
2(1− cos θ∗), (4.33a)
u = −2p1 · q2 = −s
2(1 + cos θ∗), (4.33b)
t2 + u2
s2=
1
2(1 + cos2 θ∗). (4.33c)
53
-
31. Using the following definite integral ∫ 1−1dx
1
2(1 + x2) =
4
3, (4.34)
show that ∫ 1−1d cos θ∗
t2 + u2
s2=
4
3. (4.35)
32. Substituting above results to the cross section formula, show that
σ(e+e− → µ+µ−) = 12s× 2(4π)
2α2
16π
∫ 1−1d cos θ∗
t2 + u2
s2
=4πα2
3s. (4.36a)
4.3 e+e− → q + q̄
33. Neglecting masses, show that
σ(e+e− → qq̄) = Nc × e2q × σ(e+e− → µ+µ−), (4.37)
where eq is the fractional charge of the quark.
4.4 qq̄ → g∗ → q′ + q̄′
34. Neglectin
top related